Discussions For High-precision GPS Data Processing On TBC Software

With more and more extensive application of Global Positioning System, the techniques and methodsof high-precision GPS data processing have been steadily progressing, especially in high-precision GPSdata processing software. In the field of large-scale, high-grade precision positioning, crustal deformationanalysis etc, the data processing use high-precision GPS research software. Generally, for thehigh-precision GPS control network which mainly is used in engineering, the data processing usehigh-precision GPS commercial software. In a large number of GPS commercial software, Trimble’s forsurveying and mapping project of the new generation after treatment-Trimble Business Center (TBC),is notonly high accuracy, speedy, and could process the data of GPS and GLONASS. However, the onlydrawback is differences which performance the output-reports with our GPS measurement specification andmeasurement practices, and these differences bring some trouble in the actual work.This paper use TBC data processing software as a carrier, and propose solving-solutions for commonproblems in engineering practice, enrich practical techniques and methods of the high-precision GPS dataprocessing; and develop TBC statistical software for GPS data post-processing, solve the details of theproject inconsistent with our specification. The main innovation of this paper.1. Precise ephemeris is used in a commercial GPS data processing software of baseline treatment. Byanalysis of the two currently used ephemeris interpolation method for the remaining items, and withexamples, discuss the sliding interpolation algorithm than the superiority of the non-sliding interpolationalgorithm and the best interpolation order of two interpolation algorithm; with an engineering exampleexplore how to use the precise ephemeris is loaded in TBC, propose the specific solutions andimplementation processes.2. RBF neural network model which based on the AIC criterion is established. The problems whichthe initial cluster centers, the initial expansion constant and the hidden nodes of RBF neural network modelis difficult to determine in GPS height conversion are solved. With an engineering example, the reliabilityof the model is demonstrated by comparison with the conventional model analysis.3. Data post-processing statistical software based on TBC is developed. Against the defects of TBCwhich could not output baseline processing report and GNSS closed loop report that is not accordance withthe relevant norms of GPS measurements for a comprehensive statistical analysis and quality check, andthe need for manual editing, automation weak, this paper develop TBC data post-processing softwarewhich statistical analysis from single baseline, repeat baseline, extract simultaneous loop and independentloop, the error of closed loop, GPS height conversion.The results are in line with GPS measurementstandards and measurement practices of china.